> nice to know the risks of making each of these mistakes ... only found a few places where a mistake can have real negative consequences ... context is king
That's an excellent question. As an analogy, I would say that it's similar (but not identical) to making mistakes in natural-language spelling and grammar.
For many cases, you are correct that the meaning can be repaired with context. dis iz da kase 2 wif nglish, as u kan sea w/ this sntnce.
Likewise, there will be some cases where a single letter or word can make all the difference. "He hit her" conveys a different meaning than "He hit on her"; an ESL student might only learn the verb and not the important subsequent preposition. And then there are pronunciations: https://www.youtube.com/watch?v=3Q44_A4NzjI .
Finally, by excusing the writer of all their mistakes, it puts the onus on the reader to grasp the right meaning. And if causes the reader to be confused or misunderstand, the reader has to go back and ask questions. That's a selfish shifting of effort.
> email me a 15mb (millibar) document
It looks like the symbol for bar (unit) is "bar", not "b". It is not part of SI. https://en.wikipedia.org/wiki/Bar_(unit)
You might not appreciate this, but there are fields where mb meaning millibit actually matters. Look at information theory such as data compression and error correction codes. We might say that JPEG encodes each RGB pixel with 3.0 bits, and WebP uses 2.8 bits/pixel, so WebP is 200 millibits per pixel more efficient than JPEG. There are many codec competitions where improvements are in the range of millibits per symbol (though they're written rather normally as 0.123 bits to not scare people).
By "autocorrecting" "mb" to "megabyte(s)", we lose out on the ability to express "millibit(s)".
> literally nobody speaks in terms of gigameters or teragrams or whatever
So why do we use megabytes, gigahertz, megapascals (for some industrial chemical processes), megawatts (power plants)? It's just a matter of habit. We collectively decided that kilometre is the biggest unit we'll tolerate, even though there is nothing technically wrong with megametre, gigametre, etc. We have the same problem with avoiding megagrams, kilolitres, kiloseconds, etc.
> It's roughly 1000km from my home town to the coast, so if if the moon is roughly 400,000 km away, then that's 400x the distance, pretty far.
That's nice, and it is true that when drawing attention to a comparison, the same unit should be used (Pat Naughtin and The Metric Maven both advocate for this idea). But it's also okay to write that the moon is 400 Mm away in the context of astronomy, because very few distances between bodies in space are less than 1 Mm.
The big problem with how writers treat the kilometre is that we always end up with things like "the nearest star is 4.35 light-years or 40 trillion km away". Instead of using SI prefixes as designed, we end up jamming big number words in front of the unit, and that doesn't make things any clearer. Would you like it if I said my CPU is 3.5 billion hertz, and ban all the mega-, giga-, etc.? I could even make the case that this isn't theoretical, because for example my monitor runs at 60 Hz, and we certainly must not call it 0.000 000 06 GHz with any sense of normalcy.
> My car weighs roughly 1 ton, so Hafthor's 501kg deadlift record is half my car
And you contradicted yourself because now you're using different units. It would be better to say that your car is 1000 kg and the deadlift record is 501 kg. So it's like sometimes you accept changing units, sometimes you don't.